SVM-based Classification Research Articles

An ontology-driven, SVM approach for hyperspectral image classification

ABSTRACT​​Although support vector machine (SVM) has shown powerful capability to support supervised classification for hyperspectral data, such model does not have the capability to define semantically the classification and reasoning rules which can potentially contribute to a more accurate supervised classification. In this article, we present an ontology-driven framework to support SVM-based hyperspectral data classification. First, we proposed a dimension reduction algorithm to allow automatic selection of prominent spectral characteristics for each land cover class included in a hyperspectral image. The result of the prominent spectral characteristics includes ranking and weights, used to signify the importance of a subset of all wavebands in distinguishing a particular land cover class from others. Then, we developed an ontology named HIC-Ontology to formally represent the extracted spectral characteristics to support the final training and classification process. The experimental results show that the proposed technique achieves better performance in classifying hyperspectral data than using the classic classification algorithm alone. We expect this work to contribute significantly in hyper-spectral image processing by introducing this knowledge-based approach.

DisArticle: a web server for SVM-based discrimination of articles on traditional medicine

BackgroundMuch research has been done in Northeast Asia to show the efficacy of traditional medicine. While MEDLINE contains many biomedical articles including those on traditional medicine, it does not categorize those articles by specific research area. The aim of this study was to provide a method that searches for articles only on traditional medicine in Northeast Asia, including traditional Chinese medicine, from among the articles in MEDLINE.ResultsThis research established an SVM-based classifier model to identify articles on traditional medicine. The TAK + HM classifier, trained with the features of title, abstract, keywords, herbal data, and MeSH, has a precision of 0.954 and a recall of 0.902. In particular, the feature of herbal data significantly increased the performance of the classifier. By using the TAK + HM classifier, a total of about 108,000 articles were discriminated as articles on traditional medicine from among all articles in MEDLINE. We also built a web server called DisArticle (http://informatics.kiom.re.kr/disarticle), in which users can search for the articles and obtain statistical data.ConclusionsBecause much evidence-based research on traditional medicine has been published in recent years, it has become necessary to search for articles on traditional medicine exclusively in literature databases. DisArticle can help users to search for and analyze the research trends in traditional medicine.

SVM-based sentiment classification: a comparative study against state-of-the-art classifiers

SVM-based sentiment classification: a comparative study against state-of-the-art classifiers

SVM-based sentiment classification: a comparative study against state-of-the-art classifiers

Transforming the unstructured textual information contained in various social media streams into useful business knowledge is an extremely difficult computational task, mainly, due to the underlying hard pattern classification problem of sentiment analysis, especially within the context of the Greek language. In this paper, we address the pattern classification problem of sentiment analysis through the utilisation of support vector machines (SVMs). In particular, we conducted an extensive experimental comparison where we tested the aforementioned classifier against a set of state-of-the-art machine learning classifiers on a benchmark dataset originating from the Greek bank sector by collecting data from the streaming API of Twitter that were explicitly referring to the major banks of Greece. Our results present classification accuracy and execution time metrics for each classifier, revealing the superiority of the SVM learning paradigm in assigning patterns to the correct sentiment class.

Area Determination of Diabetic Foot Ulcer Images Using a Cascaded Two-Stage SVM-Based Classification.

The standard chronic wound assessment method based on visual examination is potentially inaccurate and also represents a significant clinical workload. Hence, computer-based systems providing quantitative wound assessment may be valuable for accurately monitoring wound healing status, with the wound area the best suited for automated analysis. Here, we present a novel approach, using support vector machines (SVM) to determine the wound boundaries on foot ulcer images captured with an image capture box, which provides controlled lighting and range. After superpixel segmentation, a cascaded two-stage classifier operates as follows: in the first stage, a set of k binary SVM classifiers are trained and applied to different subsets of the entire training images dataset, and incorrectly classified instances are collected. In the second stage, another binary SVM classifier is trained on the incorrectly classified set. We extracted various color and texture descriptors from superpixels that are used as input for each stage in the classifier training. Specifically, color and bag-of-word representations of local dense scale invariant feature transformation features are descriptors for ruling out irrelevant regions, and color and wavelet-based features are descriptors for distinguishing healthy tissue from wound regions. Finally, the detected wound boundary is refined by applying the conditional random field method. We have implemented the wound classification on a Nexus 5 smartphone platform, except for training which was done offline. Results are compared with other classifiers and show that our approach provides high global performance rates (average sensitivity = 73.3%, specificity = 94.6%) and is sufficiently efficient for a smartphone-based image analysis.

Noninvasive diagnosis of chronic kidney diseases using urinary proteome analysis.

In spite of its invasive nature and risks, kidney biopsy is currently required for precise diagnosis of many chronic kidney diseases (CKDs). Here, we explored the hypothesis that analysis of the urinary proteome can discriminate different types of CKD irrespective of the underlying mechanism of disease. We used data from the proteome analyses of 1180 urine samples from patients with different types of CKD, generated by capillary electrophoresis coupled to mass spectrometry. A set of 706 samples served as the discovery cohort, and 474 samples were used for independent validation. For each CKD type, peptide biomarkers were defined using statistical analysis adjusted for multiple testing. Potential biomarkers of statistical significance were combined in support vector machine (SVM)-based classifiers. For seven different types of CKD, several potential urinary biomarker peptides (ranging from 116 to 619 peptides) were defined and combined into SVM-based classifiers specific for each CKD. These classifiers were validated in an independent cohort and showed good to excellent accuracy for discrimination of one CKD type from the others (area under the receiver operating characteristic curve ranged from 0.77 to 0.95). Sequence analysis of the biomarkers provided further information that may clarify the underlying pathophysiology. Our data indicate that urinary proteome analysis has the potential to identify various types of CKD defined by pathological assessment of renal biopsies and current clinical practice in general. Moreover, these approaches may provide information to model molecular changes per CKD.

Micrometastasis Detection Guidance by Whole-Slide Image Texture Analysis in Colorectal Lymph Nodes

Introduction/ Background Cancer is a disease that affects millions worldwide and accurate determination of whether lymph nodes (LNs) near the primary tumor contain metastatic foci is of critical importance for proper patient management. Histopathological evaluation is the only accepted method to make that determination. However, the current standard of care only examines a single central histological section per LN and yields an unacceptable false-negative rate. Aims To help pathologists in their examination we propose a method that extracts textural features from histopathological LN whole slide images (WSI) and then applies support vector machines (SVMs) to automatically identify regions suspicious for metastatic foci. Methods The database consisted of WSI from 44 LNs. Sections were stained with hematoxylin-eosin and examined at 20x (0.45μm resolution). Twenty-eight of the LNs were identified by an expert pathologist as positive for cancer (P), and the remaining sixteen were negative (N). This database was divided into two groups. Group 1 (15P and 5N) was used for training and Group 2 (13P and 11N) was used for testing the classification technique. For all analysis each WSI was divided into non-overlapping 1000 x 1000 pixel sub-images that will be referred to as high-power fields (HPFs). For each LN in Group 1, at least one WSI was annotated by a pathologist to identify rectangular, HPF-scale regions as locally cancerous or locally non-cancerous. From these annotated slides, 924 HPFs (462 P and 462 N) were obtained. For each of these HPFs, statistical features based on gray-level co-occurrence matrices [1] and Law’s texture energy measures [2, 3] were extracted from 9 derived images [4]. The extracted features were submitted to a sequential forward selection (SFS) method [5] to select few non-redundant features providing best class separation (cancerous vs. non-cancerous region). Combinations of the selected features were tested on the 924 HPFs using k-fold cross-validation to find those that produced the best results and consequently to train our SVM-based classifier. In Group 2, WSI were not annotated for cancerous and non-cancerous zones on a HPF scale. Each LN, however, had been labeled by a pathologist as positive or negative for cancer. For each WSI, each section was divided into contiguous HPFs, and those which mainly contain fatty tissue, background, and tears were automatically excluded. Each selected HPFs was classified as cancerous or non-cancerous using the previously trained classifier to obtain the total number of cancer-classified per LN. A receiver operating characteristics (ROC) curve was traced by changing the discriminator threshold (T) used to label the LN as P for cancer as a function of the total number of cancer-classified HPFs. Results During training, 5 Laws features were selected by SFS. Highly satisfactory k-fold cross-validation with a F-score of 0.996 ± 0.005 was obtained using only 2 statistical features computed at different scales. The ROC curve obtained by applying the SVM-classifier to the test set is shown in the next figure. Two valuable operating points can be identified which both guaranteed no false-negative. At T=11 we got 2 false-positives and an optimal F-score of 0.917, and with a more conservative approach, T=1, we got 7 false-positives and a F-score of 0.759. The top-left part of the slide displayed in next figure would have been proposed to the pathologist as the most suspicious region of the cancerous LN.

Successful classification of cocaine dependence using brain imaging: a generalizable machine learning approach.

BackgroundNeuroimaging studies have yielded significant advances in the understanding of neural processes relevant to the development and persistence of addiction. However, these advances have not explored extensively for diagnostic accuracy in human subjects. The aim of this study was to develop a statistical approach, using a machine learning framework, to correctly classify brain images of cocaine-dependent participants and healthy controls. In this study, a framework suitable for educing potential brain regions that differed between the two groups was developed and implemented. Single Photon Emission Computerized Tomography (SPECT) images obtained during rest or a saline infusion in three cohorts of 2–4 week abstinent cocaine-dependent participants (n = 93) and healthy controls (n = 69) were used to develop a classification model. An information theoretic-based feature selection algorithm was first conducted to reduce the number of voxels. A density-based clustering algorithm was then used to form spatially connected voxel clouds in three-dimensional space. A statistical classifier, Support Vectors Machine (SVM), was then used for participant classification. Statistically insignificant voxels of spatially connected brain regions were removed iteratively and classification accuracy was reported through the iterations.ResultsThe voxel-based analysis identified 1,500 spatially connected voxels in 30 distinct clusters after a grid search in SVM parameters. Participants were successfully classified with 0.88 and 0.89 F-measure accuracies in 10-fold cross validation (10xCV) and leave-one-out (LOO) approaches, respectively. Sensitivity and specificity were 0.90 and 0.89 for LOO; 0.83 and 0.83 for 10xCV. Many of the 30 selected clusters are highly relevant to the addictive process, including regions relevant to cognitive control, default mode network related self-referential thought, behavioral inhibition, and contextual memories. Relative hyperactivity and hypoactivity of regional cerebral blood flow in brain regions in cocaine-dependent participants are presented with corresponding level of significance.ConclusionsThe SVM-based approach successfully classified cocaine-dependent and healthy control participants using voxels selected with information theoretic-based and statistical methods from participants’ SPECT data. The regions found in this study align with brain regions reported in the literature. These findings support the future use of brain imaging and SVM-based classifier in the diagnosis of substance use disorders and furthering an understanding of their underlying pathology.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-016-1218-z) contains supplementary material, which is available to authorized users.

Multiple kernel learning based on three discriminant features for a P300 speller BCI

In this paper, we propose multiple kernel learning (MKL) based on three discriminant features to learn an efficient P300 classifier to improve the accuracy of character recognition in a P300 speller BCI. Three discriminant features are specified in raw samples and two morphological features, which can complement the MKL of a P300 classification. A linear kernel is established for each discriminant feature. A kernel weight differentiates the linear kernel to both explore complementary information among the three discriminant features and weigh a contribution of each discriminant feature for the MKL. Here, the ℓ1 norm regularization of the kernel weight ultimately enforces an optimal discriminant feature set of the MKL of a P300 classification.The performance of the proposed method is then evaluated according to the size of the three discriminant feature sets that are generated from dataset II of BCI competition III. Compared to an existing SVM-based classification method, the proposed method consistently obtains better or similar accuracy in terms of character recognition, with a different execution time for the variable size of the three discriminant feature sets. Furthermore, the kernel weight of the raw samples was found to consistently be more dominant than the kernel weight of the two morphological features on the variable size of the three discriminant feature sets. This finding means that the two morphological features supplement the lack of the raw samples for the MKL of a P300 classification. We ultimately could conclude that the proposed method is sufficiently robust to improve the accuracy of character recognition with a different time for the variable size of the three discriminant feature sets in a P300 speller BCI.

High-field MR spectroscopy in the multiparametric MRI evaluation of breast lesions

IntroductionThe aim of this work was to assess the role of 3T-MR spectroscopy (MRS) in the multi-parametric MRI evaluation of breast lesions, using a pattern-recognition based classification method. Methods291 patients (301 lesions, median 2.3cm3) were enrolled in the study (age 18–85y, mean 54.2y). T1-TSE (TR/TE=400/10ms) and T2-STIR imaging (TR/TE=5000/60ms), dynamic-contrast-enhanced MRI (DCE-MRI), apparent diffusion coefficient (ADC) (b=0–800s/mm2), and single-voxel MRS (10×10×10mm3, PRESS, TR/TE=3000ms/135ms) were performed by means of a 3T scanner. MRS results were accepted if the FWHM of the water peak was ⩽45Hz. Total choline (tCho) was considered detected if the signal-to-noise ratio (SNR) of the 3.2ppmpeak was ⩾2. A classifier-based analysis (support-vector-machines, SVM) was performed with 4-dimensional vectors including type of margin, DCE-MRI kinetic curve type, ADC mean value, and tCho SNR. A comparison with 3-dimensional vectors (without tCho SNR) was used to assess MRS impact on sensitivity, specificity, and positive-negative predictive values (PPV-NPV) for malignancy. Results228 lesions (180 malignant/48 benign) showed acceptable spectral quality. Comparison of classification results with histopathological examination of surgical specimens showed sensitivity=93.7%, specificity=84.9%, PPV=95.2%, NPV=81.5% without the inclusion of MRS in the SVM analysis. When MRS was included, the figures increased to 95.1%, 90.7%, 97.2%, and 85.0%, respectively. ConclusionsInclusion of 3T-MRS in the multi-parametric MRI evaluation of breast lesions improved the performance of the SVM-based classifier, showing a possible role of high-field MR spectroscopy in the differential diagnosis between benign and malignant breast lesions. Further research is however needed to confirm this initial evidence.

A Multiple Kernel Learning approach for human behavioral task classification using STN-LFP signal.

Deep Brain Stimulation (DBS) has gained increasing attention as an effective method to mitigate Parkinson's disease (PD) disorders. Existing DBS systems are open-loop such that the system parameters are not adjusted automatically based on patient's behavior. Classification of human behavior is an important step in the design of the next generation of DBS systems that are closed-loop. This paper presents a classification approach to recognize such behavioral tasks using the subthalamic nucleus (STN) Local Field Potential (LFP) signals. In our approach, we use the time-frequency representation (spectrogram) of the raw LFP signals recorded from left and right STNs as the feature vectors. Then these features are combined together via Support Vector Machines (SVM) with Multiple Kernel Learning (MKL) formulation. The MKL-based classification method is utilized to classify different tasks: button press, mouth movement, speech, and arm movement. Our experiments show that the lp-norm MKL significantly outperforms single kernel SVM-based classifiers in classifying behavioral tasks of five subjects even using signals acquired with a low sampling rate of 10 Hz. This leads to a lower computational cost.

Enhancement of a text-independent speaker verification system by using feature combination and parallel structure classifiers

Speaker verification (SV) systems involve mainly two individual stages: feature extraction and classification. In this paper, we explore these two modules with the aim of improving the performance of a speaker verification system under noisy conditions. On the one hand, the choice of the most appropriate acoustic features is a crucial factor for performing robust speaker verification. The acoustic parameters used in the proposed system are: Mel Frequency Cepstral Coefficients, their first and second derivatives (Deltas and Delta–Deltas), Bark Frequency Cepstral Coefficients, Perceptual Linear Predictive, and Relative Spectral Transform Perceptual Linear Predictive. In this paper, a complete comparison of different combinations of the previous features is discussed. On the other hand, the major weakness of a conventional support vector machine (SVM) classifier is the use of generic traditional kernel functions to compute the distances among data points. However, the kernel function of an SVM has great influence on its performance. In this work, we propose the combination of two SVM-based classifiers with different kernel functions: linear kernel and Gaussian radial basis function kernel with a logistic regression classifier. The combination is carried out by means of a parallel structure approach, in which different voting rules to take the final decision are considered. Results show that significant improvement in the performance of the SV system is achieved by using the combined features with the combined classifiers either with clean speech or in the presence of noise. Finally, to enhance the system more in noisy environments, the inclusion of the multiband noise removal technique as a preprocessing stage is proposed.

Computer-assisted detection of swallowing difficulty

To evaluate classification performance of a support vector machine (SVM) classifier for diagnosing swallowing difficulty based on the hyoid movement data attained from videofluoroscopic swallowing study, the hyoid kinematics during the swallowing of 2 mL of liquid barium solution were analyzed for 90 healthy volunteers and 116 dysphagic stroke patients. SVM was used to classify the kinematic results as normal or dysfunctional swallowing. Various kernel functions and kernel parameters were used for optimization. Features were selected to find an optimal feature subset and to minimize redundancy. Accuracy, sensitivity, specificity, and area under a receiving operating characteristic curve (AUC) were used to assess the discrimination performance. In 19 out of 26 features, mean comparison revealed a significant difference between healthy subjects and dysphagic patients. By reducing the number of features to 10, an AUC of 0.9269 could be reached. Common features showing the best classification in both kernel functions included forward maximum excursion time, upward maximum excursion time, maximum excursion length, upward maximum velocity time, upward maximum acceleration time, maximum acceleration, maximum acceleration time, and mean acceleration. SVM-based classification method with the use of kernel functions showed an outstanding (AUC of 0.9269) discrimination performance for either healthy or dysphagic hyoid movement during swallowing. We expect that this classification method will be useful as an adjunct diagnostic tool by providing automatic detection of swallowing dysfunction as well as a research tool providing deeper understanding of pathophysiology.

Dataflow object detection system for FPGA‐based smart camera

Embedded computer vision based smart systems raise challenging issues in many research fields, including real-time vision processing, communication protocols or distributed algorithms. The amount of data generated by cameras using high resolution image sensors requires powerful computing systems to be processed at digital video frame rates. Consequently, the design of efficient and flexible smart cameras, with on-board processing capabilities, has become a key issue for the expansion of smart vision systems relying on decentralised processing at the image sensor node level. In this context, field programmable gate arrays (FPGA)-based platforms, supporting massive data parallelism, offer large opportunities to match real-time processing constraints compared with platforms based on general purpose processors. In this study, the authors describe the implementation, on such a platform, of a configurable object detection application, reformulated according to the dataflow model of computation. The application relies on the computation of the histogram of oriented gradients and a linear SVM-based classification. It is described using the CAPH programming language, allowing efficient hardware descriptions to be generated automatically from high level dataflow specifications without prior knowledge of hardware description languages such as VHDL or Verilog. Results show that the performance of the generated code does not suffer from a significant overhead compared with handwritten HDL code.

Incremental SVM-based classification in dynamic streaming networks

Incremental SVM-based classification in dynamic streaming networks

Vehicle detection based on visual saliency and deep sparse convolution hierarchical model

Traditional vehicle detection algorithms use traverse search based vehicle candidate generation and hand crafted based classifier training for vehicle candidate verification. These types of methods generally have high processing times and low vehicle detection performance. To address this issue, a visual saliency and deep sparse convolution hierarchical model based vehicle detection algorithm is proposed. A visual saliency calculation is firstly used to generate a small vehicle candidate area. The vehicle candidate sub images are then loaded into a sparse deep convolution hierarchical model with an SVM-based classifier to perform the final detection. The experimental results demonstrate that the proposed method is with 94.81% correct rate and 0.78% false detection rate on the existing datasets and the real road pictures captured by our group, which outperforms the existing state-of-the-art algorithms. More importantly, high discriminative multi-scale features are generated by deep sparse convolution network which has broad application prospects in target recognition in the field of intelligent vehicle.

Support Tensor Machines for Classification of Hyperspectral Remote Sensing Imagery

In recent years, the support vector machines (SVMs) have been very successful in remote sensing image classification, particularly when dealing with high-dimensional data and limited training samples. Nevertheless, the vector-based feature alignment of the SVM can lead to an information loss in representation of hyperspectral images, which intrinsically have a tensor-based data structure. In this paper, a new multiclass support tensor machine (STM) is specifically developed for hyperspectral image classification. Our newly proposed STM processes the hyperspectral image as a data cube and then identifies the information classes in tensor space. The multiclass STM is developed from a set of binary STM classifiers using the one-against-one parallel strategy. As a part of our tensor-based processing chain, a multilinear principal component analysis (MPCA) is used for preprocessing, in order to reduce the tensorial data redundancy and, at the same time, preserve the tensorial structure information in sparse and high-order subspaces. As a result, the contributions of this work are twofold: a new multiclass STM model for hyperspectral image classification is developed, and a tensorial image interpretation framework is constructed, which provides a system consisting of tensor-based feature representation, feature extraction, and classification. Experiments with four hyperspectral data sets, covering agricultural and urban areas, are conducted to validate the effectiveness of the proposed framework. Our experimental results show that the proposed STM and MPCA-STM can achieve better results than traditional SVM-based classifiers.

MRI texture features as biomarkers to predict MGMT methylation status in glioblastomas.

Purpose:Imaging biomarker research focuses on discovering relationships between radiological features and histological findings. In glioblastoma patients, methylation of the O6-methylguanine methyltransferase (MGMT) gene promoter is positively correlated with an increased effectiveness of current standard of care. In this paper, the authors investigate texture features as potential imaging biomarkers for capturing the MGMT methylation status of glioblastoma multiforme (GBM) tumors when combined with supervised classification schemes.Methods:A retrospective study of 155 GBM patients with known MGMT methylation status was conducted. Co-occurrence and run length texture features were calculated, and both support vector machines (SVMs) and random forest classifiers were used to predict MGMT methylation status.Results:The best classification system (an SVM-based classifier) had a maximum area under the receiver-operating characteristic (ROC) curve of 0.85 (95% CI: 0.78–0.91) using four texture features (correlation, energy, entropy, and local intensity) originating from the T2-weighted images, yielding at the optimal threshold of the ROC curve, a sensitivity of 0.803 and a specificity of 0.813.Conclusions:Results show that supervised machine learning of MRI texture features can predict MGMT methylation status in preoperative GBM tumors, thus providing a new noninvasive imaging biomarker.

SVM-based soft classification of urban tree species using very high-spatial resolution remote-sensing imagery

ABSTRACTClassification from remote-sensing data has proved to be fast and effective for mapping tree species. However, the interference from background noise and the weak spectral separability among most tree species have negative impacts on classification accuracy and may make conventional classification methods inappropriate. To solve this problem, a soft classification approach is presented. It consists of the technologies of soft partition, defuzzifying, and feature-location analyses (referred to here as F-L analyses). The soft partition is conducted by combined binary support vector machines, enabling a natural prototype of fuzzy clusters to be formed. Unsure members (members means both patches and pixels) in the clustering prototype can be defuzzified afterwards. F-L analyses run through the entire process in two ways: (1) by adding density description to feature space for classification and (2) through density-constrained defuzzifying. The former can enrich the feature space, thus making the soft partition perform better. By using the latter, the credibility of a class label is indicated by several weighted fuzzy measures in the form of hierarchical densities, thereby reducing the uncertainty of the defuzzifying. Experiments indicate that only 63% of members on average are surely allocated in a clustering prototype and the remainder can be ensured by the explored defuzzifying approach. The mean overall accuracy (OA) of the defuzzified members is very close to that of the originally allocated sure ones. A comparative assessment shows that the mean OA and mean kappa statistic value of members defuzzified using the explored fuzzy measure are, on average, 3.84 and 4.89% respectively higher than those using conventional maximum membership measures. Experiments also show that the former is superior to the latter in regard to the formation of complete tree crown objects.

Supervised approach for detecting average over popular items attack in collaborative recommender systems

Recent research has shown the significant vulnerabilities of collaborative recommender systems in the face of profile injection attacks, in which malicious users insert fake profiles into the rating database in order to bias the system's output. To reduce this risk, a number of approaches have been proposed to detect such attacks. Although the existing detection approaches can detect the standard type of these attacks effectively, they perform badly when detecting the recently proposed obfuscated type of these attacks, for example, average over popular items (AoP) attack. With this problem in mind, in this study the author propose a supervised approach to detect such attack. First, he uses the theory of term frequency inverse document frequency (TFIDF) to extract the features of AoP attack. Second, he uses the training set to train support vector machine (SVM) to generate a SVM-based classifier. Finally, he uses the generated classifier to detect the AoP attack. The experimental results on MovieLens dataset show that the proposed approach can detect AoP attack with high recall and precision.